Wavelet Transform Processing and Classification applied to Digital Mammography

Back Back towards the Home Page

PhD Project

The aim of this project is the implementation of an image processing and pattern recognition package based on the wavelet transform and applied to mammographic images. The image enhancement procedure is based on the alteration of the multiscale wavelet transform coefficients. The pattern recognition approach is based on statistical methods and incorporates twenty image features and five classifiers. Our preliminary results suggest that: I) the W.T. enhancement technique can be used as a potential discriminator for masses abnormalities, ii) the automated classification between normal - abnormal images gives high scores of accuracy.
* Return to Research Page

* Return to Home Page

The information on this page is supplied and compiled by John Gardener
E-mail: mmooney@medphys.ucl.ac.uk
Last update 4/2/97